twilio-actions
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data for phone numbers, message bodies, and call purposes, which are interpolated into curl commands. This creates a surface for indirect prompt injection if the user provides content intended to manipulate the behavior of the outbound call agent or the downstream recipient.\n
- Ingestion points: User-provided values for 'to', 'body', and 'purpose' parameters defined in SKILL.md.\n
- Boundary markers: The skill uses JSON formatting within the curl command payload, but provides no specific instructions for escaping special characters or sanitizing content.\n
- Capability inventory: Uses the bash tool to execute curl commands for local network requests to http://localhost:8000.\n
- Sanitization: No explicit sanitization or input validation logic is described in the skill instructions.\n- [COMMAND_EXECUTION]: The skill relies on the bash tool to execute curl commands. While directed at a local proxy, this pattern grants the agent shell access to interact with network services based on potentially unvalidated user input.
Audit Metadata